While we have looked at technical analysis as a form of identifying changing trends, there is a new movement towards AI and machine learning. There is some concern that this could eventually lead to self-fulfilling prophecies, with systems following other systems, but is this really the case?


AI fuelled trading


Before we look at AI in depth, it is crucial to recognise the role of machine learning, which many see as a subset of AI. If we take a top-down approach, AI is the concept of creating smart, intelligent machines which can create a raft of recommendations or auto-transact. Machine learning is how artificial intelligence systems can learn by processing vast amounts of data. Technical algorithms will draw out patterns, responses and actions associated with human traders, which can then be replicated in an AI service.


Many high-tech dealing platforms provide historical and real-time information on trading patterns, market movements and how the two interact. As a result, we are at a stage today where AI trading systems can actually predict, with unerring accuracy, the subsequent reactions/actions of a human trader.


Piggybacking successful traders


While there still needs to be a degree of control with investors, one area which has gained significant traction in recent times is piggybacking successful traders. Numerous trading platforms use historical and real-time data to identify successful trading strategies. These are then incorporated with AI services, which will, as a consequence of machine learning, calculate the chances of individual trading strategies remaining successful.


Over time, using dealing data, market data and complex algorithms, it is possible to identify strategies which will work in specific market conditions. This is then incorporated with an automatic fund allocation system whereby individuals place funds on deposit, which are "allocated" to those traders with the best chance of continued success. Of course, even in the world of AI/machine learning, nothing is guaranteed, but the fact that these systems can even contemplate likely human reactions in particular scenarios is frightening.


Are the machines taking over?


No! However, there is no doubt that dealing systems are now more advanced and latency much lower. While the historical trading patterns of yesteryear are still there, due to the fast-moving nature of new high-tech services, the duration of these trading cycles has been reduced. With the best will in the world, there will never be an AI system which can predict human nature with 100% accuracy. However, what about the flip side?


Could we arrive at a situation where traders become over-reliant on AI trading systems, and they effectively become self-fulfilling prophecies? At this moment in time, this is unlikely because the authorities are constantly battling to reduce the influence of these automated systems. It would just take one system to malfunction due to corrupt code or incorrect data, and markets could go into freefall. As similar systems followed the trend, momentum would build, and markets would be effectively at the beck and call of machine-led trading.




In theory, some of the systems available today can predict how the human brain thinks with unerring precision (not 100% accuracy). As they constantly learn, the margin of error will reduce, and in theory, the influence of these AI systems will grow. We are now at a stage where we depend on regulators regulating - it is as simple as that!

Back to News